PROJECT SUMMARY (See instructions); Abstract Olfactory information is encoded by combinations of olfactory bulb glomeruli, which individually represent distinct chemical features of different odorants. This information is transmitted to the piriform cortex, where it drives activation of ensembles of neurons whose concerted activity is thought to lead to a holistic odor percept. Piriform neurons are also interconnected by a rich recurrent circuitry. Here, we propose to extend our previous work dissecting the cellular mechanisms that shape these corticalensembles. We are particularly interested in the problem of concentration invariance. Odorant receptor specificity depends on ligand concentration, with more receptors being activated at higher odorant concentrations. However, odors retain their identity over a large range of concentrations, requiring that odor representations be normalized with respect to concentration somewhere along the olfactory pathway. Such normalization occurs in neural circuits throughout the brain, although the cellular mechanisms underlying this normalization remain poorly understood. Recent experiments suggest odor-evoked activity in piriform cortex is relatively concentration invariant, consistent with the relatively concentration-invariant quality of the odor percept. In the mentored phase of this award we examined the recurrent piriform network in vitro and discovered features of this circuit that are poised to play a major role in shaping odor-evoked cortical activity and suggest a circuit-level solution to effecting concentration invariance. Here, we will quantitatively characterize the transformation of odor representations from olfactory bulb to piriform cortex using in vivo two-photon microscopy to image odor-evoked activity in populations of olfactory bulb mitral cells and principal neurons in piriform cortex (Aim 1). We will build an experimentally constrained, network-level computational model to test the sufficiency of our circuit-level model for effecting concentration invariance (Aim 2). Finally, we will use genetic and viral strategies to selectively eliminate the recurrent circuitry to directly test the role of these circuits in mediating concentration invariance (Aim 3). This research program employs a rich arsenal of conceptual and technical approaches that have